December 5, 2019

Voice is the next frontier.

I write this text in 2019, and it's almost 2020. Growing up in the 80s, I looked forward to the year 2000 and how things would be different in a future dominated by robots. And yet during those years, the year 2000 was just 19, 18 or 15 years away.
Now we are 20 years past the year 2000, and technology evolved in a way that each person turned into an island, a robot like figure. An android.
The granularity in society is so small today, mainly thanks to the way smartphones have concentrated so much functionality of our lives. It's like we are extensions of our own phones instead of the other way around. They provide music, video, news, snippets into strangers lives, allow me to sell and buy digital assets, write, talk and see people from the other side of the world, monitor my baby, be warned when to put the bins outside, pay for things, read books, listen to books & podcasts, make and edit videos, photos and write texts like these, store my loyalty cards, pay for car parking based on my location, do an insurance claim, identify myself in a secure way (replacing ID Cards), check in for a flight, order food, buy a train ticket, see if it's going to rain, be guided while driving, make a doctor's appointment, check when my package is arriving through post, ... the list is almost infinite.

Does it even make sense to still call it phones?
With this rectangle of glass and metal you can travel the world and if this device is connected, you can buy an airplane ticket, rent a room in Bologna, Santiago, Brisbane or Lisbon; order food for your wife in Amsterdam while you are sitting on an airport in Seattle; or order a taxi when you got lost walking through Vilnius. Smartphones have been delivering society changing innovation, but are still pretty much treated as edge devices, except for video and photography. In fact, there is not a lot of creation and building happening on smartphones if you exclude social media content and video streaming.

We all know they can do more. Way more. Much more. But do we have the tools today to do it?

So what about the next 20 years? What about the 20s, and 30s of this century? Will "phones" even be the smart device that we would interact more with?

Let's step back and understand the trends:

1) Phones will remain for many years to come, will become more powerful, but the cloud computing systems that back their apps is where the magic will happen. This means, phones will still be a sort of edge device between each person and the internet.

2) Voice interface is growing but we have the problem that language is not the best interface. I mean, some very simple tasks take a lot of wording in most languages. So context will be key for voice.

3) Context. This is where phone and voice interface come as one. If I say the word "tea" whilst in the kitchen part of my house, that could very well be the keyword to put my kettle on. But if I say "tea" inside my car, nothing happens, or an answer could pop up: "you are too far from home so for security reasons the kettle will not be turned on". But if I say "Bach", regardless where I am, a selection of music composed by Johann Sebastian Bach can actually play inside the car, in the kitchen, or on my phone if I am in public transport using my bluetooth earphones.

Whoever wins the voice interface technology battle will be the provider of the (cloud) backend and interface API to make this a reality. Until that happens we have already lots of voice innovations, but they all are trying to mimic human voice communication skills, mainly language. Innovations like Amazon Alexa and Amazon Polly, Google Assistant or even Apple Siri.
Where voice can play a massive role in the future technology interface systems, is if used to transmit instructions in a non human fashion and very much dependant on context.

It's like a married couple, or a pair of very good friends who when faced with a 3rd person speaking something at them, just look at each other and immediately establish a communication and information transmission just with a facial expression and sometimes not even that: that situation in that particular context triggers in both listeners the same thought process or memory of a common experience.

Coded voice communication is what I am talking about and this is what will come once machines learn how to interpret context and even complex thoughts transmitted through voice.

But what is all this good for you may ask? This will be a class of innovation at the abstraction layer that will glue all the things that scientists and technology engineers are working today, at the cognitive systems foundational level. This means that I can go out to walk the dog, or just walk, have a series of ideas and thoughts and very quickly my edge interface (aka "phone") is able not only to capture them, but to act on them, without me having to articulate them in the language of humans.

Does this mean we need a new language for voice to become more ubiquitous as a technology interface? Maybe. Man develops new programming languages every few years to pass on instructions to computers, why not create another one that is able to encompass the same complexity but using voice instead of lines of code?

Who wants to have a go?

August 14, 2019

Dropping the mic

This is not a tech post. Sorry. In fact my relationship with my blog has been the most open possible. I've started blogging in 2007, mainly to put out notes from my notebook about how I solved customer problems and to document the tools I developed in that process.
When I joined Oracle Corporation in 2009, I had to either delete the blog, or delete every single post where I had URLs to documents that were exclusive to customers who paid support. That was the right thing to do, so in order to keep my blog open, I deleted all those posts where I would say that if you have problem A you should go and read document B from Oracle MySupport website. It was not even ethical to keep it online, since those correlations where based on my own experiences, but someone else with a similar problem, may have needed a deeper root cause analysis so it was not right to keep those pieces of writing up online.
When I joined Oracle I became more technological fluent, but less technical, so my blogging style changed. Again I could not even blog 10% of the material I was creating internally at Oracle, because I was lucky enough to be working constantly in new areas, and most of the stuff was still not public. So again I stopped blogging altogether.
Now I left Oracle, and joined Amazon. At Amazon you are more encouraged to write, but I'm still representing a company and that writing has a specific context.

So before I know what to do with this blog, I come back every now and then, and publish something.

Well, this time I will share the article I wrote 2 weeks ago, which explains my last career move. Hope you enjoy.

https://www.linkedin.com/feed/update/urn:li:article:7926540364894830085/

November 2, 2018

There's no technology without science, there's no science without data

In today's world, technology is far more popular than science, and this is an astonishing fact taking into consideration that one cannot exist without the other. It's not a chicken and egg situation of course, but in most young consumers of technology, and even some of the most savvy ones, technology seems to have displaced science of its "cool" status. Let's look at some stats.


Well, from the above we can see "Science" is typically more searched for, than "Technology". But this is not a good metric for two reasons:
1) People do not search for "Scientific" themes by putting the "Science" word on it. A quick look at scientific trends will reveal some specific trends like "Biotechnology", or "Global Warming".
2) For the same reasons as above people do not search for "Technology" but more the trends. Technology is so pervasive these days that we really need to split between the consumer technology and the enterprise technology. In consumer technology the smartphone is king and the obsession is still very much present in the market, so let's go with the two more popular models: "Samsung Galaxy" and "Apple iPhone". In enterprise computing the two main trends are "Blockchain" and "Artificial Intelligence".

So let's see how some of these trends stack against each other in Google Trends. First the two science trends we identified above: "Biotechnology" versus "Global Warming".


Very close! Global Warming had a sudden rise in the last month, but all in all very much comparable. So we can choose either one to stack against the technology trends.
Now let's look at both technology trends classes. First consumer technology.


Not a surprise that the smartphone market only has spikes whenever there is a new model out! In the case of the Samsumg Galaxy that spike is in February and in the case of the Apple iPhone that happens in September. Now let's see how the enterprise technology trends compare against each other.


I just realised that "Artificial Intelligence" is not comparable with "Blockchain", because the first one just captures searches in English language. As you can see below "Blockchain" is present across the world, while "Artificial Intelligence" is very language dependent.


"Global Warming" is of course very language dependent. So we need to compare things that are either language neutral or in the same language. And that means comparing "Global Warming" from the science side with "Artificial Intelligence" from the technology side, and then "Biotechnology" from the with "Blockchain". And finally stack up the "winners" from both against each other and finally end this long rant about how Science is not as popular as Technology.

Science wins!

Technology wins!


Technology wins again :)

Don't lose sight of the title of this post. The fact that technology is a result of scientific discovery and technology is applied science is not clear to everyone, in particular the people who depend on technology, which is most of us in the developed world. I say "us" because if you are reading this you have to own a device to access the internet and that puts you right on the half of the world's population who has internet access. Which in itself is also a technology wonder. And that is precisely the point with today's wonder and awe: it's all directed at technology and very few at science. But science in order to move forward needs to use technology in order to process its data. Let's recap what is the "scientific method" and why it relies so much on data it captures from observation.

Related image
There are two important aspects to the scientific method that rely on its source data: data quality, and ability to store the intermediate results from all the experiments. Field science, applied science or even non-theoretical scientists have to come up with ways to speed up the process of testing hypothesis, which in turn will lead to other hypothesis, and this work can take years. It's a complex tree of ramifications of hypothesis and at each new level more data is generated, and the more ramifications and branches it creates, the bigger the number of combinations and the harder it gets to combine and process all this data. But can they all afford to store, analyse and process all that data? In most cases scientists and researchers work with simulation tools that will allow them to speed up the process of analysing the data generated in real time, or at the same time as the data is being generated.
So the frontier for the scientific research that will push technology to the next level is dependent on the data that science is able to process today, which in turn dependes on the technology of today. Unless we have a breakthrough in technology so big, that it will expand the limits of science research.